Sentence-Level Search System for Viewpoint-Based Article Retrieval
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current search systems using morphological analysis fail to efficiently collect information related to a specific viewpoint or partial content in an article, often returning articles with similar sentences but no new content, or articles with different content altogether.
Innovation Solution
A search system that receives a specific sentence as input, searches for articles with similar meanings, and displays related articles in descending order of difference relative to the original article, using entailment relationships and clustering to identify relevant content.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If morphological analysis is used to search for articles containing similar sentences, then the search system can find articles with similar content, but it cannot efficiently collect information related to a specific viewpoint and may return articles with no new content
Solution Approach 1:
The patent segments an article into multiple sentences and enables users to specify a particular sentence as the search target. This segmentation allows the search to focus on a specific viewpoint or partial content within an article, rather than treating the entire article as a single unit. By dividing the article into sentence-level units, the system can precisely target the user's information need while filtering out unrelated content.
Solution Approach 2:
The patent applies local quality by performing morphological analysis and entailment determination specifically on the specified sentence rather than the entire article. This localized processing ensures that the search captures articles containing similar viewpoints or connotations while maintaining efficiency. The system analyzes the semantic properties of the target sentence in isolation, then uses those properties to search for related articles.
2Stability of the object's composition
If the search system returns articles with similar sentences, then it can maintain consistency with the user's query, but it may return articles that are completely different from the user's viewpoint of interest
Solution Approach 1:
The patent introduces entailment determination as an intermediary mechanism between morphological analysis and search result generation. After performing morphological analysis on the specified sentence, the system uses entailment determination to assess whether candidate articles truly convey the same viewpoint or connotation. This intermediary step filters out articles that merely contain similar words but differ in meaning, ensuring both consistency and accuracy in search results.
Solution Approach 2:
The patent replaces traditional keyword-based mechanical matching with semantic-based entailment determination. Instead of relying solely on word overlap or simple similarity metrics, the system uses natural language processing to understand the meaning and logical relationships in sentences. This substitution transforms the search mechanism from a mechanical pattern-matching system to a semantic understanding system, improving viewpoint accuracy.
3Productivity
If the search system uses traditional morphological analysis, then the processing is simple and fast, but it cannot effectively identify articles with new content about the user's viewpoint
Solution Approach 1:
The patent performs preliminary morphological analysis and entailment determination on the user's specified sentence before conducting the search. This preliminary processing extracts key semantic features and establishes a robust search query representation. By preparing the search criteria in advance through these NLP techniques, the system improves search efficiency while ensuring that articles with new content related to the user's viewpoint are accurately identified.
Solution Approach 2:
The patent changes the search parameters from simple keyword matching to semantic features derived from morphological analysis and entailment determination. Instead of searching based on exact word matches, the system searches based on extracted semantic properties such as part-of-speech tags, lemmatized forms, and entailment relationships. This parameter transformation enables the system to efficiently identify articles with new content that semantically align with the user's viewpoint.
Data Source
AI summary
A search system includes a reception unit 601 for receiving specification of a specific sentence in an article, a search unit 602 using a first sentence being the specified sentence as a search query to search an article including a sentence having a meaning similar to the connotations of the first sentence, and a display unit 603 for displaying information of the searched article as a related article to the first sentence, and when there are a plurality of related articles, the display unit 603 preferentially displays the related articles in descending order of difference relative to a first article being the article including the first sentence.